I hope I get help here. I am still in the early stages of statistical analysis with R.
Anyway, here is the problem:
Study design:
20 patients assessed at 3 different time points: Timepoint 0 = Baseline ; Timepoint 20 (20 days with Treatment A/Placebo respectively), Timepoint 40 (40 days after Baseline ; 20 days with Treatment A/Placebo respectively). We can safely assume in this setting that there is no soever carry over effect. Sequence was randomized (10 Patients started with Treatment A => Placebo, 10 Patients Placebo => Treatment A)
At the three time points the same 4 assessment tests were performed (Test_1; Test_2; Test_3; Test_4) (all continuous measurements)
While I understand that I can test the different groups with paired T T-Test / Wilcoxon signed-rank test depending on distribution I am not sure how to asses the influence of the Baseline performance of the individuals, the sequence assignment, Sex or e.g. disease duration and previous treatment history (coded as 1, 2 or 3)
I thought about using a linear mixed-effects model fit (NLME package) with the following command:
Test1 <- lme(Test_1 ~ Group:Group + SeqTx + Age + factor(Sex) + factor(prevTx) + yearsofdiseases, random=~1|StudyID, data = dataset)
Test2 <- lme(Test_2 ~ Group:Group + SeqTx + Age + factor(Sex) + factor(prevTx) + yearsofdiseases, random=~1|StudyID, data = dataset)
The Dataframe looks as follow Dataframe, and it's columns are:
StudyID: Participant ID
Group: Timepoint 0 (Baseline), 20 days after Treatment A (Treatment A), 20days after Placebo (Placebo)
SeqTx: 0 = (Place=>TreatA), 1 = (TreatA => Placebo)
prevTx: prev TX factorial (1,2 or 3)
Is the approach right? How do I interpret if a variable influences outcome. As far As I understand the result at the moment tells what changes for expample if I take the GroupPlacebo the Test1 increases comapred to Baseline. But e.g. how do I get how Baseline influences the Test on Placebo or TreatA
Sorry I hope it is not too confusingly written but i am a little desperate....
